Lessons from application of the UNRES force field to predictions of structures of CASP10 targets

Abstract

The performance of the physics-based protocol, whose main component is the United Residue (UNRES) physics-based coarse-grained force field, developed in our laboratory for the prediction of protein structure from amino acid sequence, is illustrated. Candidate models are selected, based on probabilities of the conformational families determined by multiplexed replica-exchange simulations, from the 10th Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction (CASP10). For target T0663, classified as a new fold, which consists of two α + β domains homologous to those of known proteins, UNRES predicted the correct symmetry of packing, in which the domains are rotated with respect to each other by 180°in the experimental structure. By contrast, models obtained by knowledge-based methods, in which each domain is modeled very accurately but not rotated, resulted in incorrect packing. Two UNRES models of this target were featured by the assessors. Correct domain packing was also predicted by UNRES for the homologous target T0644, which has a similar structure to that of T0663, except that the two domains are not rotated. Predictions for two other targets, T0668 and T0684_D2, are among the best ones by global distance test score. These results suggest that our physicsbased method has substantial predictive power. In particular, it has the ability to predict domain-domain orientations, which is a significant advance in the state of the art.protein folding | structure symmetry | multi-domain packing P rediction of protein structures from amino acid sequence still remains an unsolved problem of computational biology. Although, since the famous experiments by Anfinsen (1), it is known that a protein adopts the structure which is the (kinetically reachable) global minimum of the free energy of a system, it is not straightforward to implement this physical principle in practice because of the inaccuracy of existing force fields and because of the enormous difficulty to search the conformational space of the system. Therefore, the most effective methods for protein-structure prediction nowadays are knowledge-based approaches, in which database information is incorporated explicitly into the procedure (2). These methods can be divided into three categories, namely, comparative (homology) modeling (3-5), in which the target sequence is compared with the sequences for which experimental structures are known and those structures are usually selected as candidate models for which the greatest similarity is observed; threading (6-8), in which the target sequence is superposed on structures from a database, and those which give the highest score (lowest pseudoenergy) are selected as candidate predictions; and, finally, the fragment-assembly or minithreading method developed by David Baker and colleagues (9, 10), in which the predicted structure is assembled from nine-residue fragments extracted from a protein-structure database, and knowledge-and physicsbased filters are applied at each asse...

Performance of protein-structure predictions with the physics-based UNRES force field in CASP11

“…The procedure used by us in CASP11 differs by several aspects from those used in CASP10 (He et al, 2013) and in previous CASP experiments (Liwo et al, 1999;Ołdziej et al, 2005). First, owing to improved parallelization and code optimization performed lately, as well as access to larger supercomputer resources, we were able to treat all CASP11 targets released for human prediction instead of selecting only those which were judged not to be TBM targets.…”

“…Our physicsbased approach to protein-structure prediction, that uses the UNRES force field, performed reasonably well in biannual CASP Community Wide Experiment on the Critical Assessment of Techniques for Protein Structure Prediction, exercises to obtain fold topology, e.g. for targets T0061, T0063 and T0079 in CASP3 (Lee et al, 1999(Lee et al, , 2000, T0102 in CASP4 (Liwo et al, 2011), T0129 andT0149 in CASP5, T0215, T0223, T0230 andT0281 in CASP6 (Ołdziej et al, 2005), T0534, T0537 and T0578 in CASP9 (Liwo et al, 2011) and T0644, T0663, T0668 and T0740 in CASP10 (He et al, 2013;Khoury et al, 2014). However, because the UNRES predictions are generally of medium-resolution quality, which is remarkably lower than those for TBM targets, UNRES was featured only for new-fold targets (proteins with unique orientation of the local structure) and for targets with new types of domain packing.…”

“…In the CASP10 experiment, two UNRES-based predictions made only by the Cornell-Gdansk group, and by one other group with additional participants, were featured by the assessors (He et al, 2013). The first one was a two-domain target T0663 (treated by the Cornell-Gdansk group) (He et al, 2013), and the second one was T0740 (treated by the wfCPUNK group within the WeFold initiative) (Khoury et al, 2014); in the second prediction exercise, UNRES was supplemented with contact-prediction restraints.…”

“…The first one was a two-domain target T0663 (treated by the Cornell-Gdansk group) (He et al, 2013), and the second one was T0740 (treated by the wfCPUNK group within the WeFold initiative) (Khoury et al, 2014); in the second prediction exercise, UNRES was supplemented with contact-prediction restraints. In the CASP10 experiment, we proved that the UNRES force field is a very good tool to predict the orientation of the domains for the 2-fold symmetry target T0663 (He et al, 2013) that was further confirmed by post-CASP tests with use of restraints within the domains, but not between them (Krupa et al, 2015).…”

Prediction of Protein Structure by Template-Based Modeling Combined with the UNRES Force Field

“…This result was achieved through correct prediction of the structures of individual domains, even though the domain packing was incorrect, as opposed to that obtained with UNRES. 9 Because the UNRES models of this target had the correct domain packing and UNRES supplemented with contact prediction information resulted in the best prediction of T0740, 10 in this work a combined approach is proposed in which knowledge-based modeling is used for those parts of the protein sequence for which the bioinformatics signals are strong and physics-based modeling with UNRES is implemented to obtain the correct domain packing. It is highly probable that the new approach can bridge the gap between the bioinformatics methods, which are accurate when good templates can be found but can fail when all or part of the target sequence exhibits too-weak similarity to those of database structures, and the physics-based methods, which are more robust because of the independence of structural databases but are more expensive and less accurate for template-based modeling (TBM) targets than the bioinformatics approaches.…”